Depth Conversion Methods for the Torsk Oilfield: Investigating the Complex Velocity Field of the Seaspray Group, Gippsland Basin
Honours Degree, 2003
University of Adelaide
The accuracy of depth conversion in the Gippsland Basin is a critical factor in the success of exploration and development projects. Severe channelling and subsequent infill in the section above the reservoir have created a complex velocity field over the basin. Sediments deposited within the channels are up to 2000 ms-1 faster than the surrounding strata.
Interpretation of seismic and well data revealed several sizeable channels over the Torsk structure within the Seaspray Group. The velocity effects of these channels may have resulted in the Torsk-1 well being drilled off crest. Several depth conversion techniques were employed to estimate the depth to the primary reservoir in the Gippsland Basin (the Top Coarse Clastics horizon). The resulting maps were analysed to compare the merits of the depth conversion techniques, and to examine possible locations for a second well up dip from Torsk-1.
Seven depth maps were created for analysis using various types of velocities. Average velocities were estimated from both well data and from seismic stacking velocities. Interval velocities were calculated from stacking velocities using the Dix equation.
A major difficulty in the analysis of these depth conversions was the lack of well control with which to test the resulting depth maps. Hence, less direct methods such as the predicted hydrocarbon column height at the wells and the variation of the predicted depths between maps were used. This was aided by the calculation of an average depth map, the differences between the individual maps and the average map, and standard deviation and percentage difference in the maps.
Analysis of the depth conversion techniques suggested that they did not, in fact, account for all the complexities of the velocity field in the area. The differences in depth were all below 3 percent. However, this equates to a depth difference of over 30 m and is a sizeable error in a structure that has an average relief of 35 m. The lowest standard deviations occurred, as expected, close to the well locations. Despite this, critical points such as the crest and spill point of the structure were well defined laterally. Thus, a better location for a well up dip from Torsk-1 well could be determined. However, the additional column height is poorly defined and further investigation is recommended to reduce risks associated with depth conversions.